Underdetermined blind separation based on potential function with estimated parameter's decreasing sequence
FU Wei-hon · Systems engineering and electronics · 2014
A potential function method with estimated parameter is proposed for blindness of scale parameter selecting of k-dimensional subspace method.Firstly,the reference value r1is estimated based on observed signals and the scale parameterσ1for potential function f is estimated based on r1.When cluster planes are estimated by f,the reference value r2is estimated based on them and the scale parameterσ2for potential function g is estimated based on r2.Thus,we obtain the estimation of mixing matrix.This method makes full use of the statistical properties of observed signals.The experiment shows that this method is reduced by 75%in matrix estimation error than the method before improved and is reduced by 1-2orders of magnitude compared with other methods when there is no noise;the estimated precision is 8.7%higher and the accuracy rate of the source' s number estimation is twice higher.